[zeromq-dev] Multicore Magic
sustrik at 250bpm.com
Wed Apr 28 12:48:30 CEST 2010
Brian Granger wrote:
> BUT, the other part is scalability and this is where I am still
> hopeful. Maybe we will be stuck with horrible latency in Python
> because of this issue. But if that latency scales well (is constant)
> with the total number of processes, we will be OK. Traditional
> threading with shared memory and locks, etc. is super fast....until
> you have hundreds or thousands of threads all chomping on the locks.
A horrible hack is still possible: run several python interpreters in
parallel in the same process.
To make it work python interpreter cannot use any global variables.
That's probably not the case.
Anyhow, it's a nasty hack.
More information about the zeromq-dev